FECUNDITY OF NORTH AMERICAN SALMONIDAE 



453 



that Titcomb's data are of limited value since 

 Titcomb stated that some of the charr had ap- 

 parently dropped part of their eggs before being 

 captured. Osgood Smith (1947) obtained a linear 

 relation between the logarithm of tlie number of 

 eggs and the body length of 29 eastern charr, 

 but inasmuch as his specimens were from such 

 diverse localities as California, Ontario, and 

 Xortli Carolina, the results cannot be regarded as 

 conclusive. 



The number of eggs of eastern charr from four 

 localities is shown according to size of fish in table 

 1 and figure 1. These data show that the dif- 

 ferences in egg number between localities are too 

 great to permit combining localities in studying 

 the egg number-fish size relation. When the 

 curves for the separate localities are examined, it 

 becomes apparent that the number of eggs in- 

 creases appro.ximately as the weight of the fish, 

 since the logarithm of egg number plotted against 

 the logarithm of fish length approximates a straight 

 line, as does the logarithm of fish weight against 

 the logarithm of fish length. 



Table 1. — Fecundily of eastern charr, Salvelinus fontinalis 



2 500 



1000 



CO 



o 



LlI 



li. 

 o 

 a: 



LlI 



m 



500 



250 



100 



-\ 1 1 — I — r 



-n/ew JERSEY 



(HATCHERYI 



20 25 30 40 50 60 70 80 100 



FORK LENGTH (CM.) 



Figure 1. — Relation of egg number to body length in 

 Salvelinus fontinalis. 



Locality and fork lenplh 



W voniing (beaver pond) : 

 12.70 cm 



14.44 cm 



l(i.82cm 



19.65cm - 



New Jersey (hatchery stock): 



29.69 cm ' 



31.00 cm 



32.1.1cm 



35.21 cm 



36.50 cm 



38.00 cm - 



Michigan (streams): 



10.40 cm » 



12.T3cm 



15.07 cm 



17.41 cm , 



19.74 cm _ 



22.08 cm 



24.42 cm - 



Quebec (Laiireiitides Park): 



13.75 cm 



16.25 cm 



18.75 cm 



21.25 cm 



23.75 cm 



28.25 cm --- 



28.75 cm.. 



31.45 cm 



35.50 em 



Authority 



Allen (1956). 



Hayford and Em- 

 body (1930). 



Cooper (1953). 



Vladykov and 

 Leeendre (1940). 



1 Converted from standard length by (actor 1.1. 

 :Converted from total length by factor 0.92. 



The logarithm of the number of eggs shows a 

 closer linear relation to length than does the 

 actual number of eggs when specimens are avail- 

 able over a wide range of length. However, 

 over the rather narrow ranges of length at maturity 

 found in Oncorhynchus, the difference is usually 

 trifling and can be disregarded in computing. 



Extensive data on the relations between number 

 of eggs and length and weight of the fish are given 

 by Foerster and Pritchard (1941). Correlations 

 between egg number and fish length and between 

 egg number and fish weight are shown for Cultus 

 Lake sockeye for each of 6 years (1932-35, 1937, 

 and 1938) and for pijik salmon from McClinton 

 Creek, Masset Inlet, for eich of 6 even-numbered 

 years from 1930 to 1940, inclusive. 



In order to compare the values of the two series 

 of correlations, we have combined the correlation 

 coefficients for the various years by transforming 

 the r values into z values (Fisher 19:10, p. 171). 

 The value of /• for the combined samples is ob- 

 tained from the weighted average value of z. 

 The results are as follows: 



V'aiut of correlaticn of 



egg number with — 

 FM length Fitit weight 



0. 57 0. 56 



0. 35 0. 40 



Sockeye salmon (Cultus Lake).. 

 Pink salmon (McClinton Creek). 



This means that in the sockeye about 31 to 32 

 percent of the variation in number of eggs is 



